Assignment 5 - Spatial Data Exploration

UCID : 12149052

Name : Regina Widjaya

CLOROPLETH MAP 1

## OGR data source with driver: ESRI Shapefile 
## Source: "ne_50m_admin_1_states_provinces_lakes", layer: "ne_50m_admin_1_states_provinces_lakes"
## with 100 features
## It has 40 fields

## [1] "SpatialPolygonsDataFrame"
## attr(,"package")
## [1] "sp"

## [1] "+proj=longlat +datum=WGS84 +no_defs +ellps=WGS84 +towgs84=0,0,0"

m <- ggplot(h1b_state, aes(long, lat, group=group, fill= perc_diff)) + 
  #geom_text(aes(long, lat, group = NULL, label = h1b_state$h1b_1000), size = 2) +
  geom_polygon() +
  #geom_text(aes(long, lat, label = scales::percent(h1b_state$perc_diff)), color = "white") +
  coord_map("albers", lat0 = 29.5, lat1 = 45.5,
            xlim = c(-124.85, -66.88), ylim = c(24.4, 49.38),
            orientation = c(90, 0, -98.35)) +
  labs(title="H1-B median wage is being paid MUCH MORE than host state median wage...", 
       subtitle="Percent Difference in Median Wage (Local State vs. H1-B)", 
       caption="Source: USCIS", 
       y="Latitude",
       x= "Longitude") +
  guides(fill=guide_legend(title="--% Higher")) +
  scale_fill_viridis("perc_diff", labels= scales::percent, direction = -1) +
  geom_path(color="white") + 
  theme(legend.position = "bottom",
        axis.title.x= element_text(family = 'Courier New', face = 'bold', size = 10),
        axis.text.x = element_text(angle = 0, hjust = 1, family = 'Andale Mono', size = 8),
        axis.title.y = element_text(family = 'Courier New', face = 'bold', size = 10),
        axis.text.y = element_text(angle = 0, hjust = 1, family = 'Andale Mono', size = 8),
        panel.background=element_rect(fill="#f7f7f7"),
        plot.background=element_rect(fill="#f7f7f7"),
        strip.background = element_rect(fill="#f7f7f7"),
        legend.background = element_rect(fill="#f7f7f7"),
        plot.title = element_text(family = 'Courier New', hjust = 0, face = "bold", size = 15),
        plot.subtitle = element_text(family = 'Courier New', face = "italic", size = 10),
        plot.caption = element_text(family = 'Courier New', face = "italic", size = 9),
        panel.grid.major.x=element_line(color="gray100"),
        panel.grid.minor.x=element_line(color="gray90")
        ) 

plot(m)

CLOROPLETH MAP 2

m2 <- ggplot(h1b_state, aes(long, lat, group=group, fill= perc_diff_CM)) + 
  #geom_text(aes(long, lat, group = NULL, label = h1b_state$h1b_1000), size = 2) +
  geom_polygon() +
  #geom_text(aes(long, lat, label = scales::percent(h1b_state$perc_diff)), color = "white") +
  coord_map("albers", lat0 = 29.5, lat1 = 45.5,
            xlim = c(-124.85, -66.88), ylim = c(24.4, 49.38),
            orientation = c(90, 0, -98.35)) +
  labs(title="However, H1-B wage are still less than \nmedian Computer & Mathematics wage...", 
       subtitle="Percent Difference in Median Wage (Local State vs. H1-B)", 
       caption="Source: USCIS", 
       y="Latitude",
       x= "Longitude") +
  guides(fill=guide_legend(title="--% Higher")) +
  scale_fill_viridis("perc_diff", option = "magma", labels= scales::percent) +
  geom_path(color="white") + 
  theme(legend.position = "bottom",
        axis.title.x= element_text(family = 'Courier New', face = 'bold', size = 10),
        axis.text.x = element_text(angle = 0, hjust = 1, family = 'Andale Mono', size = 8),
        axis.title.y = element_text(family = 'Courier New', face = 'bold', size = 10),
        axis.text.y = element_text(angle = 0, hjust = 1, family = 'Andale Mono', size = 8),
        panel.background=element_rect(fill="#f7f7f7"),
        plot.background=element_rect(fill="#f7f7f7"),
        strip.background = element_rect(fill="#f7f7f7"),
        legend.background = element_rect(fill="#f7f7f7"),
        plot.title = element_text(family = 'Courier New', hjust = 0, face = "bold", size = 15),
        plot.subtitle = element_text(family = 'Courier New', face = "italic", size = 10),
        plot.caption = element_text(family = 'Courier New', face = "italic", size = 9),
        panel.grid.major.x=element_line(color="gray100"),
        panel.grid.minor.x=element_line(color="gray90")
  ) 

plot(m2)

INTERACTIVE MAP (PLOTLY)

intr_df <- h1b_state[,c("name","total_h1b","median_wage","h1b_1000","perc_diff","perc_diff_CM", "pop_company", "postal")]
intr_df <-unique(intr_df)
intr_df$perc_diff <- scales::percent(intr_df$perc_diff)
intr_df$perc_diff_CM <- scales::percent(intr_df$perc_diff_CM)


intr_df$hover <- with(intr_df, paste(name, '<br>', 
                          "# of Approved H1-B", total_h1b, '<br>', 
                          "Median Wage", median_wage, "<br>",
                          "*", perc_diff,"Above State Median Wage","<br>",
                          "*", perc_diff_CM,"Above State C&M Wage","<br>",
                          "Biggest Sponsor:", pop_company))

# give state boundaries a white border
l <- list(color = toRGB("white"), width = 2)
# specify some map projection/options
g <- list(
  scope = 'usa',
  projection = list(type = 'albers usa'),
  showlakes = TRUE,
  lakecolor = toRGB('white')
)

pl <- plot_ly(intr_df, z = intr_df$total_h1b, text = intr_df$hover, locations = intr_df$postal, type = 'choropleth',
        locationmode = 'USA-states', color = intr_df$total_h1b, colors = 'Purples',
        marker = list(line = l), colorbar = list(title = "# of Approved H1-B Application")) %>%
  layout(title = '2016 H1-B Approved Application Summary<br>(Hover for breakdown | Source: USCIS)',
         caption = 'Source: USCIS' , geo = g)

pl